mathiascreutz commited on
Commit
e110f3a
1 Parent(s): d6a4efb

Minor modifications

Browse files
Files changed (1) hide show
  1. README.md +15 -14
README.md CHANGED
@@ -155,7 +155,9 @@ data = load_dataset("GEM/opusparcus", "de.100")
155
  data = load_dataset("GEM/opusparcus", "fr.90")
156
  ```
157
 
158
- TODO: Add comment about larger and noisier sets being better for training.
 
 
159
 
160
  ### Data Instances
161
 
@@ -246,9 +248,8 @@ up in the datasets.
246
  The training sets were not annotated manually. This is indicated by
247
  the value 0.0 in the `annot_score` field.
248
 
249
- For an assessment of of inter-annotator agreement, see Mikko Aulamo,
250
- Mathias Creutz and Eetu Sjöblom (2019). [Annotation of subtitle
251
- paraphrases using a new web
252
  tool.](http://ceur-ws.org/Vol-2364/3_paper.pdf) In Proceedings of the
253
  Digital Humanities in the Nordic Countries 4th Conference],
254
  Copenhagen, Denmark.
@@ -312,16 +313,16 @@ approximately 1000 sentence pairs that have been verified to be
312
  acceptable paraphrases by two indepedent annotators.
313
 
314
  The `annot_score` field reflects the judgments made by the annotators.
315
- If ´the annnotators fully agreed on the category (4.0: dark green,
316
- 3.0: light green, 2.0: yellow, 1.0: red), the value of
317
- `annot_score` is 4.0, 3.0, 2.0 or 1.0. If the two annotators
318
- chose adjacent categories, the value in this field will be 3.5, 2.5 or
319
- 1.5. For instance, a value of 2.5 means that one annotator gave a
320
- score of 3 ("mostly good"), indicating a possible paraphrase pair,
321
- whereas the other annotator scored this as a 2 ("mostly bad"), that
322
- is, unlikely to be a paraphrase pair. If the annotators disagreed by
323
- more than one category, the sentence pair was discarded and won't show
324
- up in the datasets.
325
 
326
  #### Who are the annotators?
327
 
 
155
  data = load_dataset("GEM/opusparcus", "fr.90")
156
  ```
157
 
158
+ Remark regarding the optimal choice of training set qualities:
159
+ Previous work suggests that a larger and noisier set is better than a
160
+ smaller and clean set. See Sjöblom et al. (2018). [Paraphrase Detection on Noisy Subtitles in Six Languages](http://noisy-text.github.io/2018/pdf/W-NUT20189.pdf). In Proceedings of the 2018 EMNLP Workshop W-NUT: The 4th Workshop on Noisy User-generated Text., and Vahtola et al. (2021). [Coping with Noisy Training Data Labels in Paraphrase Detection](https://aclanthology.org/2021.wnut-1.32/). In Proceedings of the 7th Workshop on Noisy User-generated Text.
161
 
162
  ### Data Instances
163
 
 
248
  The training sets were not annotated manually. This is indicated by
249
  the value 0.0 in the `annot_score` field.
250
 
251
+ For an assessment of of inter-annotator agreement, see Aulamo et
252
+ al. (2019). [Annotation of subtitle paraphrases using a new web
 
253
  tool.](http://ceur-ws.org/Vol-2364/3_paper.pdf) In Proceedings of the
254
  Digital Humanities in the Nordic Countries 4th Conference],
255
  Copenhagen, Denmark.
 
313
  acceptable paraphrases by two indepedent annotators.
314
 
315
  The `annot_score` field reflects the judgments made by the annotators.
316
+ If the annnotators fully agreed on the category (4.0: dark green, 3.0:
317
+ light green, 2.0: yellow, 1.0: red), the value of `annot_score` is
318
+ 4.0, 3.0, 2.0 or 1.0. If the annotators chose adjacent categories,
319
+ the value in this field will be 3.5, 2.5 or 1.5. For instance, a
320
+ value of 2.5 means that one annotator gave a score of 3 ("mostly
321
+ good"), indicating a possible paraphrase pair, whereas the other
322
+ annotator scored this as a 2 ("mostly bad"), that is, unlikely to be a
323
+ paraphrase pair. If the annotators disagreed by more than one
324
+ category, the sentence pair was discarded and won't show up in the
325
+ datasets.
326
 
327
  #### Who are the annotators?
328